Editorial

Abstract

Introduction

Software

Software Evaluations

Assessment

Summary

Conclusions

Acknowledgements

References

Case Studies Index

6. Conclusions and Recommendations

It is clear that no single package provides the full range of tools identified here as being desirable in an ESDA package, in a truly interactive, graphical environment. From the evidence of this limited review, each of the two approaches to software development used in the packages reviewed has some merit, and could be used to develop a more comprehensive system:

• Through the extension of a free-standing package such as cdv or MANET. The challenge here would be to add further facilities such as spatial analytical tools and the ability to examine the implications of modifying the spatial framework.
• Through the development of a link with GIS software as with SAGE and SpaceStat. Here the challenge is to provide a more interactive and flexible graphical environment.

The facilities such a system should contain can be identified, drawing on the examples of best practice from each of the four packages reviewed in this report:

• The provision of a wide range of cartographical and statistical views of data in linked windows.
• A range of different tools for selecting cases in any given window. For example, in a map window, the ability to select one or more areas, all areas within a defined shape, all areas along a trajectory etc, with the possibility of dynamic update of other windows as the selection tool is moved.
• A `spreadsheet' type display of the non-spatial data and the ability to create new variables from exisiting ones using mathematical operators.
• The interactive modification of histogram bin sizes and symbol sizes on point symbol maps.
• The ability to turn on or off the labelling of graph axes.
• The linking of multiple cartographic views of the same data. For example, linking a cartogram with a choropleth map of the same data greatly assists in the interpretation of the cartogram.
• The ability to deal with variable base populations.
• The capability to assess the sensitivity of findings to changes in the spatial framework. Facilities here might inlcude the ability to modify the weights matrix or to produce new sets of regions.
• The provision of specialist tools for exploring spatial properties of the data, in addition to the linking of a map with graphical views.
• Handling of missing values.

Within the time available for this study it was not possible to make a very thorough evaluation of the graphical effectiveness of the software packages. However we feel that this is an important aspect of visualization software, which appears to have been overlooked by the Scientific Visualization community, and which deserves closer attention. The classification proposed by Cleveland (1994) was used as the basis for our subjective evaluation, and we feel that this may form a useful starting point for the development of a more thorough methodology in this area.
Work on visualization tools for area-based spatial data has taken place within the framework of existing statistical and cartographic software, independently of ViSC software such as Iris or AVS (Earnshaw and Wiseman 1992).This may be in part due to the way in which visualization of this type of spatial data has grown out of initial concerns with the lack of spatial analysis facilities in GIS systems. Whatever the reason, it would be extremely interesting to see whether ViSC products could in fact be used for the visualization of area-based data, since they are already available to the academic community and structures have been put in place to support them and this is also suggested as a topic for future research.

7. Acknowledgements

We are indebted to the authors of the software packages described here, who have have given freely of their time and expertise to assist us in carrying out this review: Jason Dykes (cdv), Anthony Unwin (MANET) and Luc Anselin (SpaceStat). We are aware that we have had a rather limited time in which to make our evaluation, and hope that this has not caused us to misrepresent any of the systems. We would stress again that the intention of this case study is to present a review of what is the current state of the art in the visualization of area-based spatial data across the range of available software packages, rather than making detailed comparisons between them.